Researchers at the University of Virginia have developed a method of analyzing Twitter use to predict crimes.
According to a research paper recently published in the scientific journal “Decision Support Systems,” a certain type of analysis can be applied to geo-tagged tweets to predict 19 to 25 different types of criminal activity before they even occur, including stalking, theft and multiple types of assault.
Research lead Matthew Gerber of the university’s Predictive Technology Lab said it’s obvious that no one would ever tweet about a crime they intend to commit in a direct manner, and that the key to predicting crimes through Twitter is quite the opposite.
“What people are tweeting about are their routine activities. Those routine activities take them into environments where crime is likely to happen,” Gerber said in an MSN report. “So if I tweet about getting drunk tonight, and a lot of people are talking about getting drunk, we know there are certain crimes associated with those things that produce crimes. It’s indirect.”
Gerber said that tweets without any mention of a crime can still carry a significant amount of information about activities related to crime. The team conducted its study test in Chicago, and analyzed tweets tagged in relation to certain neighborhoods sectioned off in square kilometers, and by the city’s crime database, which is already the foremost in “Minority Report”-esque criminal predictability.
Researchers were then able to formulate a glimpse into the future and make “useful predictions” about areas where certain types of crime were likely to occur, which is a huge benefit to allocating and deploying police resources to deal with incidents.
“This approach allows the analyst to rapidly visualize and identify areas with historically high crime concentrations,” the study said. “Future crimes often occur in the vicinity of past crimes, making hot-spot maps a valuable crime prediction tool.”
Gerber said that Twitter data can be easy to use because it’s publicly available and contains a large amount of tagged location data for the majority of users that never bother or care to disable such features.
“I send our algorithms to these locations and see what people are talking about,” Gerber said. “The computer algorithm learns the pattern and produces a prediction.”
The study was funded by the U.S. military, which uses similar data and analyses to foresee threats in active war zones like Iraq and Afghanistan. Gerber has already been contacted the the New York Police Department, which hopes to replicate the way the system was successfully used in Chicago. Gerber plans to extend the research into other social media areas as well.
The team’s findings were similar to other recent research into the amount of information available from Twitter, which can be data-mined, analyzed and harnessed to determine vast amounts of private, personal information about individuals and groups of users.